A Free Service from bikesport.pro
Your complete cycling platform: Track analysis & simulation, personalized training plans, bike fitting, and nutrition coaching (coming soon). Available on iOS and Android.
GPX Elevation Corrector
This tool corrects the elevation profile of GPX files using high-resolution SRTM (Shuttle Radar Topography Mission) elevation data from OpenTopography.
How It Works
1. Upload & Analyze
Upload your GPX file and the app determines which SRTM tiles (1° × 1° areas) are needed for your route.
2. Download Tiles
SRTM tiles are downloaded from OpenTopography and cached locally for future use.
3. Correct Elevations
Each track point's elevation is corrected using bilinear interpolation from the 30m resolution SRTM data.
4. Smooth & Optimize
Apply smoothing algorithms to remove grid artifacts and optionally reduce point density.
Smoothing Methods
Choose from four professional smoothing algorithms to best suit your data and use case:
| Method | Use Case | Speed | Quality | Shape Preservation |
|---|---|---|---|---|
| Moving Average | General use, default | ⚡⚡⚡ | ⭐⭐⭐ | ⭐⭐ |
| Gaussian | Smooth visualization | ⚡⚡ | ⭐⭐⭐⭐ | ⭐⭐⭐ |
| Savitzky-Golay | Mountain routes | ⚡⚡ | ⭐⭐⭐⭐⭐ | ⭐⭐⭐⭐⭐ |
| Median | Noisy GPS data | ⚡ | ⭐⭐⭐ | ⭐⭐⭐⭐ |
Best for: General purpose, everyday use
Simple centered moving average that's fast and predictable. Works well for most GPX files with moderate noise levels.
Best for: Presentation and visualization
Uses a bell curve to weight nearby points - closer points have more influence. Creates smoother, more natural-looking elevation profiles.
Best for: Mountainous terrain with peaks and valleys
Fits a polynomial to preserve important elevation features. Maintains the character of climbs and descents better than simple averaging.
Best for: GPS data with spikes and outliers
Uses the median value instead of mean, making it excellent at removing elevation spikes caused by GPS errors or bad readings.
✨ New Advanced Features
Problem: GPS artifacts create unrealistic slopes (e.g., 458% when reality is 20%)
Solution: Automatically detect and correct extreme point-to-point slopes
Three methods:
- Local Smoothing: Smooth only steep sections, preserve flat terrain
- Gradient Capping: Hard limit maximum slope (e.g., 20%)
- Smooth + Cap: Both methods for heavily corrupted data
Example: Original 458% uphill → Corrected 24.4% uphill
Problem: GPS tracks have thousands of unnecessary points (big files, slow apps)
Solution: Intelligent algorithms that preserve route quality
Two methods:
- Min Distance: Simple - keep points N meters apart
- Douglas-Peucker: Advanced - preserves curves, distance, and terrain
Result: Reduce file size by 50-90% while maintaining route accuracy
Problem: Your GPX already has good elevation data but is noisy
Solution: Skip SRTM correction, just smooth existing GPS data
Benefits:
- ⚡ Much faster (no tile download)
- ✅ Preserves original GPS measurements
- 🎯 Perfect for cleaning tracks from Garmin, Wahoo, etc.
Interactive Chart: Drag points on the elevation profile to manually correct problem areas
Features:
- 🖱️ Click and drag any point to adjust elevation
- 🔄 Automatic interpolation between edited points
- 📊 Live statistics update after changes
- ↩️ Undo/reset anytime
Processing Parameters
What it does: Controls the size of the smoothing window
Window size: 2 × radius + 1 points
Default: 3 (window of 7 points, ~540m span with typical GPS spacing)
Recommendations:
- 0-2: Minimal smoothing, preserves most detail
- 3-5: Good balance (recommended for most routes)
- 6-10: Aggressive smoothing for very noisy data
What it does: Minimum elevation change to count as gain/loss in statistics
Default: 3.0 meters
Purpose: Filters out small elevation fluctuations (noise) when calculating cumulative gain/loss
Recommendations:
- 1-3m: For smoothed data or precision requirements
- 3-5m: Standard threshold for most routes
- 6-10m: For raw SRTM or very noisy GPS data
What it does: Minimum horizontal distance for calculating slope percentages
Default: 100 meters
Purpose: Prevents unrealistic slope values from closely-spaced GPS points
Why it matters: GPS points can be very close together (1-10m), making small elevation differences appear as huge slopes. This parameter ensures slopes are measured over meaningful distances.
What it does: Secondary correction pass targeting extreme point-to-point slopes
Default Threshold: 20% (cycling), 30% (hiking)
Purpose: Fix GPS artifacts that create impossible gradients (50-400%)
Methods:
- Smoothing: Apply local smoothing only to steep sections (preserves rest of profile)
- Capping: Hard limit slope to threshold (ensures no slope exceeds X%)
- Smooth + Cap: Both methods combined (most aggressive)
Real-world example: A ride with 458% point-to-point slopes (GPS error) was corrected to 24% (realistic cycling gradient)
What it does: Intelligently reduces file size while preserving route quality
Default: None (keeps all points)
Two Algorithms:
Keeps points at least N meters apart
- 10-15m: Moderate reduction (good for most routes)
- 20-30m: Significant reduction
- 50m+: Aggressive (may lose detail on curves)
Intelligent shape-preserving algorithm - removes points based on perpendicular distance from simplified path
- 5m tolerance: Good balance (recommended)
- 10m tolerance: Significant reduction with good quality
- 15-20m: Aggressive but preserves curves better than Min Distance
Advantage: Preserves total route distance and turning points much better than Min Distance
Features
SRTM Correction
Uses 30m resolution SRTM data with bilinear interpolation for accurate elevations.
Smart Smoothing
Four professional algorithms to remove noise while preserving route characteristics.
Interactive Chart
View detailed statistics, zoom/pan elevation profile, and manually edit points by dragging.
Tile Caching
Downloaded SRTM tiles are cached locally for instant reuse on overlapping routes.
Point Reduction
Douglas-Peucker or Min Distance algorithms to reduce file size while preserving track shape.
Filtered Export
Option to export only SRTM-corrected points, excluding any uncorrected data.